Is Vehicle Plate Corner Prediction by Vision Transformer Better than CNNs?

Table 2

The number of model parameters of selected ViT backbone models along with corresponding structural parameters.

Model name	Patch size	Embedded dim	Depth	# of heads	# of param (m)
ViT-Patch-13	13	16	128	2	0.96
ViT-Patch-26	26	32	128	2	1.9
ViT-Patch-52	52	128	64	2	13.3
ViT-Patch-104	104	128	128	8	26.8
ViT-Patch-208	208	64	128	8	9.1